A 2015 Project Encoded Light Propagation Physics as a Neural Network to Reconstruct Objects
prof_kamilov · x · 2026-08-27
Ulugbek Kamilov revisits his favorite 2015 project: representing the physics of light propagation through a 3D object as a neural network.
Instead of learning the network weights, they optimized its nodes, which represented the object's unknown refractive index. The network was essentially the physics itself, and "training" it reconstructed the object — an early example of the physics-as-network idea for inverse problems.
More from Research
- Phil Engel on Claude's Proposed Complex Structure on S^6 — littmath · 2026-08-27
- End-to-End RL Drone Policy Passes Sim2real on Multiple Hardware — yacineMTB · 2026-08-27
- Agent Architecture Insight: Compressing History vs. Real-time Updates — curious_vii · 2026-08-27
- How, When and Why to Use Agentic AI in Neuroscience Labs — KordingLab · 2026-08-27
- H3 Fun ControlNet Brings Control-Video Guidance for MiniMax-H3 to ComfyUI — wyzborrero · 2026-08-27
- Prediction: Robotics Will Be Solved with RL and Simulators — tribbloid · 2026-08-27